A concise conceptual model for material data and its applications in process engineering

Abstract

Material data are of basic importance for the design, operation, and maintenance of process plants. To support computer-based modeling, calculation, and management of material data, a conceptual data model is required. Although the issue of material data modeling has been addressed in a number of efforts, a model, which is concise, self-contained, and well structured, is still not available. In this work, a systematic categorization of material data has been performed. This leads to a conceptual material data model consisting of three partial models, namely ‘substance’, ‘phase system’, and ‘phase system mathematical model’, each representing a separate perspective of material data in accordance with the physical basis as well as practical uses. This conceptual data model can be used for the development of detailed data models that are suitable for specific applications by making extensions and refinements to it. Such applications of the conceptual model are explained by the experiences gained in two software development projects.

Introduction

Material data, physical properties, and thermodynamics are of major importance in various applications in the field of chemical engineering. Process simulation and optimization as well as many design calculations and thus the entire process design are based on physical property calculations. Their results rely on the availability and quality of material and physical property data. Due to this fact, material-related information and services play a major role within process engineering application tools. Many commercial simulation tools like Aspen Plus or Hysys are provided with large pure- and multi-component databases containing a variety of pure-component and mixture property data, parameter regression tools, and services for calculating physical properties and phase equilibria. There are also extended databases holding material-related information, which are commonly used during process design. Examples are DETHERM (Westhaus, Droge, & Sass, 1999), which holds thermodynamic and transport properties of pure components and mixtures, the thermophysical property database DIPPR (Wilding, Rowley, & Oscarson, 1998), and CAMPUS, a database covering material information like rheological and mechanical properties and parameters of polymers, which are needed for the design of plastics processing facilities (CWFG, 2001). Furthermore, within different companies, large amounts of material and equilibrium data are kept in different databases. These data have been gathered over a long period of time and are now a valuable source of corporate knowledge about materials processed in a company's production facilities.

While each of these software tools and databases is quite useful on its own, there is a large potential benefit in combining these different sources and services in order to obtain integrated and consistent material data collections and physical property calculation services. The engineers should be provided with reliable and consistent material and thermodynamics information. This applies especially for open and component-based software environments, which are the promising future generation of CAPE tools (Braunschweig, Pantelides, Britt, & Sama, 2000). As such environments may support a wide range of activities, not only ‘classical’ thermodynamics but also any other material-related information that might be needed for calculations (such as plastics or electrolytes data) should be considered for integration.

For the development of these kinds of integrated facilities, it is desirable to obtain a unified view of all information that is of importance for material and physical property services. Such a unified view can be obtained in a powerful and efficient manner by the development of an appropriate information model. During the development of software systems, information modeling is a continuous activity that generates results on various levels of detail for different development stages. A conceptual model is usually considered as the starting point, which provides a common understanding of the domain of interest by describing the major concepts and their dependencies. This model is then used in the further software development activities as the basis for more detailed design and implementation models that formalize the information generated, stored, and transferred by the software systems (Fowler & Scott, 1997). In a broad perspective, such a conceptual data model essentially plays the role of an ontology of the material domain. It improves the communication between people and the inter-operability among software systems, and also brings system engineering benefits including reusability, reliability, and easiness in requirements specifications (Uschold & Gruninger, 1996). Such an ontological view to conceptual data modeling has been taken in this work.

As proposed in the literature (e.g. West & Fowler, 1996, Fox & Gruninger, 1998), a conceptual data model should meet requirements such as completeness, generality, extensibility, flexibility, perspicuity (being easy to understand), reusability, and minimality (consisting of a minimum number of necessary concepts). The evaluation of the quality of data models could be done in a rigorous manner through some quantitative approaches (e.g. Kesh, 1995, Genero, Piattini & Calero, 2000). However, these are yet not mature enough for being well accepted in the data modeling practice.

The specific requirements for a conceptual model for material data and the existing work in that area are discussed in Section 2. In Section 3, a model developed to meet these requirements is presented. This conceptual model is part of CLiP—a comprehensive data model for chemical engineering (Bayer, Krobb & Marquardt, 2001). It has been applied for the development of different applications providing and dealing with thermodynamics and material information, which will be presented in Section 4. Some concluding remarks are given in Section 5.